A new methodology, resulting from an international collaboration including the Spanish National Research Council (CSIC), promises to speed up and lower the cost of one of the most complex stages in the discovery of new medicines. This innovative approach merges high-throughput X-ray crystallography, computational tools for selecting molecular fragments, robotic systems for generating compounds, and biophysical and structural techniques to assess their binding to target proteins.
María José Sánchez-Barrena, a researcher at the Blas Cabrera Institute of Physical Chemistry (IQF) of the CSIC, highlights that the research findings address a key obstacle: efficiently leveraging structural information from interactions between proteins and small fragments. "X-ray crystallography can precisely reveal where a small fragment binds to a protein, but converting that information into more complex molecules with biological activity typically requires numerous cycles of design, synthesis, and evaluation," she explains. The new strategy connects these processes, automates part of the intensive work, and generates molecules of great chemical diversity.
The strategy has been successfully applied to a protein involved in the regulation of central nervous system processes and associated with neurological disorders: Calcium-Sensing Neuronal 1 (NCS-1). This protein, with a complex interaction surface, is a demanding pharmacological target for selectively modulating protein-protein interactions. Thanks to this method, candidate molecules have been identified to modulate interactions with neuronal receptors involved in neurodegeneration, with a promising candidate found for restoring neuronal activity in Alzheimer's disease.
The development, carried out at the XChem facilities of the Diamond Light Source Synchrotron (United Kingdom) by a multidisciplinary team led by Frank von Delft from the University of Oxford, included a predoctoral stay by Daniel Muñoz Reyes, the paper's first author. In a single 3-month experimental cycle, over 250 chemically diverse compounds were designed and synthesized from fragments identified by crystallography. Low-cost robotic synthesis and rapid evaluation, even from reaction mixtures, reduced the need for individual purification, allowing efficient exploration of chemical space and determination of binding modes to the protein.
The results demonstrate that combining high-throughput crystallography, computational tools, automated chemistry, and biophysical analysis can transform structural information into a direct driver for discovering new molecules. This connected and modular approach allows experimental data to continuously guide the next step, adapting to different proteins and projects, and expanding the exploration of molecules beyond commercially available collections.
